At the ACM International Conference on Intelligent Virtual Agents (IVA) in Berlin, Sept. 16–19, ICT is highlighting the next release of the Virtual Human Toolkit (VHToolkit)—a powerful, multi-platform testbed for creating intelligent conversational agents.
Arno Hartholt, Director of ICT’s Integrated Virtual Humans Group, presented recent advances alongside colleagues including Dr. Jonathan Gratch, Director of the Affective Computing Group, and PhD computer science student Bin Han.
The updated VHToolkit, built on ICT’s RIDE simulation environment, integrates audio-visual sensing, speech recognition, natural language processing, text-to-speech synthesis, and nonverbal behavior generation. It supports desktops, mobile devices, and AR/VR headsets, with demonstrations showcasing interoperability and the use of third-party assets.
The platform—developed by Arno Hartholt, Ed Fast, Kevin Kim, Andrew Leeds, Edwin Sookiassian, and Sharon Mozgai, Director, Virtual Human Therapeutics Lab—will be released publicly on GitHub later this fall.
Keep an eye on vhtoolkit.ict.usc.edu for the release announcement.
Thanks to Spencer Lin, XR researcher for the photos!